NRC Drought Response

This report covers the results of a geo-spatial analysis aimed to greater understand the impact of the 2018 drought in Badghis province and the Murghab catchment in north west Afghanistan. The primary aim was to identify suitable areas where Norwegian Refugee Council (NRC) can support the development and improvement of livelihoods in order that returning refugees re-establish themselves in a self-sufficient manner. With humanitarian funds from the Norwegian Ministry of Foreign Affairs.

Alcis carried out a geospatial analysis encompassing an impact assessment using Normalised Difference Vegetation Index (NDVI) – a proxy indicator for crop health – for the Area of Interest (AOI) and catchment sub basins. Alcis also carried out human impact assessment on vegetation loss and produced community profiles using livelihood potential, to locate areas of high potential livelihoods. Finally, resettlement priority areas were identified by combining the data.

The figures below show the NDVI reduction in 2018, when compared to the long-term average for the 1st crop cycle. Figure 1 illustrates the NDVI scores categorised 1 - 7 and Figure 2 shows the subtraction of the 2018 NDVI compared to the norm, categorised into significantly high - significantly low impact. Both maps show a reduction in NDVI in 2018 when compared to the long-term average. Areas that suffered greatest drought loss when compared to the long-term is the southwest and northern areas of Badghis province with the greatest reductions seen in Ghormach and Murghab districts.

In Afghanistan there are 2 cropping cycles. The same analysis has been carried out for the second cropping cycle. However, this data is less informative as it does not provide a sufficient change in data and the nature of the cropping patterns in this part of Afghanistan are such that the 2nd annual cropping cycle is normally restricted. However, in Figures 3 and 4, there is evidence for a reduction in NDVI in the northern areas of the AOI, particularly in Ghormach and Murghab districts of Badghis province.

To further understand the spatial variation in NDVI losses, a sub-basin analysis of the Murghab catchment was conducted. The results in Figure 5 illustrates the sum of the NDVI losses per sub-basin. The sub-basins with the greatest NDVI losses are the central areas of Badghis province, sub-basins C, E, and F. Sub-basin C has a total loss of -96,148. This number is generated from the 0.23 - 1 NDVI values which are reclassified 1-7 for the long term average and 2018. The difference in 2018 produces values from -6 (greatest possible losses) to +6 (greatest possible increases). The map shows the sum of the losses between -1 to -6 per sub-basin. The fewest NDVI losses are the south eastern sub-basins G, H and I. However, these are also areas of smallest agricultural area...

To take into account the area of agriculture within each sub-basin to ensure the data is proportional, Figure 6 shows the percentage of agriculture lost. Here, the spatial variation is similar, but the sub-basin with the greatest percentage of agriculture lost is sub-basin E, at 64%. This incorporates the districts Murghab, Ghormach and Qaysar. Sub-basin I has the smallest percentage of agriculture lost at only 5.2% incorporating districts Jawand and Ferozkoh. 

As part of this analysis, to support the returning refugees, it was important to understand the spread of populations throughout the AOI, to make assumptions of areas where populations are likely to have moved elsewhere. Using spot mapping from high resolution imagery Alcis has mapped all the compounds in Afghanistan. From this a 1km population density is shown below in Figure 7. The district with the highest number of compounds at 32,499 is Ferozkoh. However, for the purpose of this project, it is important to understand how many compounds are reliant on 1km2 of agriculture per district. Kohistanat has 102.72 compounds per km2 of agriculture, indicating a high compound density per km2 of ag. However, Gulran has 8.2 compounds per km2 of agriculture. 

Using the data produced thus far, Alcis generated a bivariate choropleth map (Figure 8) to assess human impact of vegetation loss. Using a 1km2 hex grid, the count of compounds and the mode of the NDVI losses in 2018 are displayed.

The increase colour darkness indicates an increase in values. For example, a darkening of pink represents an increase in population only, and a darkening of turquoise represents an increase in NDVI reduction compared to the long-term average only. Lastly, an increase in purple combines the two variables and represents an increase in both population and NDVI losses – the areas that have been hit the hardest by the drought.

Figure 8 highlights three key areas for returning refugees in the local to Qala-e-Naw, Murghab and Ghormach. To enhance this, 100k mapping was used to further locate villages in these darker purple areas. This includes: Qala-e-Naw district centre and surrounding villages: Senjetk and Dahan-i-Gharghaitu in Qala-e-Naw district, Muqur district centre Jawkar and Sini in Murghab district and Ab-i-Garmak and Sartakht in Ghormach district. All these areas lie within Badghis province.

The Graph in Figure 8 also aids decision making by comparing the area of agriculture lost per compound. Here the compounds in Muqur were most affected with a 0.054 km2 of agriculture lost per compound.

For the returning refugees, it is desirable that there is a high degree of livelihood potential in the locations of resettlement. Using a weighted sum method, combining 8 different factors that contribute to a high livelihood potential, the areas with the highest potential are shown in Figure 9. The contributing factors include: roads, makets, electricity, quality of agriculture, irrigation, population, district level data, project NSP’s, distance to district centres and distance to provinical centres. The data shows the areas with the highest livelihood potential and therefore most appropriate for returning refugees is the areas in south west Badghis - Qala-e-Naw and Murqur - the areas along the river in Murghab district and central Ghormach district. These are the areas that are purple/blue in Figure 9. 

The final province/catchment wide investigation was combining the livelihoods layer with the bivariate choropleth. The elevation of the hex’s is an extrusion from the ground of the maximum livelihood potential per hex. Therefore, from this the priority areas for resettlement works would be the confluence of high population/high or moderate impact of drought/high to moderate livelihood (purple areas with height). Below is an interactive 3D scene that displays the data. 

It is evident from the data so far that there has been a substantial reduction in agricultural production throughout Badghis province and the rest of the AOI. The irrigation channels including rivers, ditches and water lines, are vital to ensure that the rainwater that does fall, is efficiently transported around the districts. Therefore, an irrigation analysis has been conducted throughout the AOI.

As the maps state, Muqur, Ghormach, Qala-e-Naw and Qaysar districts had the least effective irrigation, and a low km of irrigation for every km2 of agriculture. This is the same for sub-basins C, E and F. For the aim of this report, and for the NRC to effectively relocate refugees this data is essential.

For returning refugees to establish themselves in a self-sufficient manner, as part of the NRC's recovery support, improvements and additional irrigation systems could be implemented in the districts most affected. This could include constructing new irrigation channels, repairing previous ones or making improvements such as dredging the channels. 

Due to the ever increasing growth of poppy cultivation in the northeast areas of Afghanistan, in particular, Badghis province, Alcis has considered the effect of the drought on poppy probability by using the previous drought year in 2008 as an indicator.

Firstly, Figure 12 shows the impact of the 2008 drought compared to the long term average. There is a significant reduction in NDVI throughout the AOI.

The maps below in Figures 13 and 14 show the poppy probability changes following the drought in 2008 - 2009/10. They clearly illustrate a substantial increase in poppy probability, predominantly in areas most affected by the drought. This is likely due to a coping strategy of growing a reliable cash crop such as poppy in order to recoup financial losses.

Figure 15 below shows the long term trend for poppy probability in the AOI. The northern areas of Badghis province have the highest long term poppy probability.

Furthermore, Figure 16 below shows the confluence of areas that experienced a reduction in poppy probability from 2007 - 2008, and a subsequent increase in 2009 and 2010. The map shows this covers a large area of the AOI, and therefore supports the thoughts that poppy cultivation exceeded, to offset the loss in 2008. This can be an indicator that poppy cultivation is going to increase in the northern districts of Badghis province in 2019 and 2020. Qades and Murghab districts are most likely.

The map in Figure 17 below shows the likely 2019/2020 poppy return areas by illustrating the confluence of the 2018 drought impact and the long term poppy probability.

It is important to consider that this may be stretching the capabilities of the data as it is only a probability dataset.

To conclude, there are 3 main areas that are most suitable for the return of refugees into Badghis. These are the districts of Qala-e-Naw, Murghab, Muqur and Ghormach. In addition to this, the report also mentions village names that are most suitable for the return of refugees. These areas are chosen as they combine the greatest reduction in NDVI, with the highest populations and greatest livelihoods potential.

Alcis has also considered the recovery potential districts using an irrigation analysis and concluded the districts which would benefit the most from development or improvements to irrigation channels. Namely, this was Muqur, Ghormach, Qala-e-Naw and Qaysar districts.

Lastly, Alcis used the 2008 drought as an indicator to assess the likely recovery in drought affected areas. The data indicated a widespread increase in poppy probability, particularly in drought affected areas, following the 2008 drought.